From 756ba4dc6986b7298aef29a8691cb00786a0f8e4 Mon Sep 17 00:00:00 2001 From: Thomas Dehaeze Date: Fri, 17 Feb 2023 11:30:40 +0100 Subject: [PATCH] Remove src folder --- matlab/uniaxial_1_micro_station_model.m | 1 - matlab/uniaxial_2_nano_hexapod_model.m | 1 - matlab/uniaxial_3_disturbances.m | 1 - matlab/uniaxial_4_dynamic_noise_budget.m | 1 - matlab/uniaxial_5_active_damping.m | 1 - matlab/uniaxial_6_hac_lac.m | 1 - nass-uniaxial-model.html | 887 +++++++++++++++++------ nass-uniaxial-model.matlab | 2 - nass-uniaxial-model.org | 2 - 9 files changed, 651 insertions(+), 246 deletions(-) diff --git a/matlab/uniaxial_1_micro_station_model.m b/matlab/uniaxial_1_micro_station_model.m index 8710abc..2aa2c5f 100644 --- a/matlab/uniaxial_1_micro_station_model.m +++ b/matlab/uniaxial_1_micro_station_model.m @@ -6,7 +6,6 @@ s = zpk('s'); %% Path for functions, data and scripts addpath('./mat/'); % Path for data -addpath('./src/'); % Path for functions %% Colors for the figures colors = colororder; diff --git a/matlab/uniaxial_2_nano_hexapod_model.m b/matlab/uniaxial_2_nano_hexapod_model.m index 32b4539..a4faf61 100644 --- a/matlab/uniaxial_2_nano_hexapod_model.m +++ b/matlab/uniaxial_2_nano_hexapod_model.m @@ -6,7 +6,6 @@ s = zpk('s'); %% Path for functions, data and scripts addpath('./mat/'); % Path for data -addpath('./src/'); % Path for functions %% Colors for the figures colors = colororder; diff --git a/matlab/uniaxial_3_disturbances.m b/matlab/uniaxial_3_disturbances.m index ea4492f..511a601 100644 --- a/matlab/uniaxial_3_disturbances.m +++ b/matlab/uniaxial_3_disturbances.m @@ -6,7 +6,6 @@ s = zpk('s'); %% Path for functions, data and scripts addpath('./mat/'); % Path for data -addpath('./src/'); % Path for functions %% Colors for the figures colors = colororder; diff --git a/matlab/uniaxial_4_dynamic_noise_budget.m b/matlab/uniaxial_4_dynamic_noise_budget.m index a7e4b85..e94fa60 100644 --- a/matlab/uniaxial_4_dynamic_noise_budget.m +++ b/matlab/uniaxial_4_dynamic_noise_budget.m @@ -6,7 +6,6 @@ s = zpk('s'); %% Path for functions, data and scripts addpath('./mat/'); % Path for data -addpath('./src/'); % Path for functions %% Colors for the figures colors = colororder; diff --git a/matlab/uniaxial_5_active_damping.m b/matlab/uniaxial_5_active_damping.m index 8ad1346..dffff45 100644 --- a/matlab/uniaxial_5_active_damping.m +++ b/matlab/uniaxial_5_active_damping.m @@ -6,7 +6,6 @@ s = zpk('s'); %% Path for functions, data and scripts addpath('./mat/'); % Path for data -addpath('./src/'); % Path for functions %% Colors for the figures colors = colororder; diff --git a/matlab/uniaxial_6_hac_lac.m b/matlab/uniaxial_6_hac_lac.m index e78bd84..2c92676 100644 --- a/matlab/uniaxial_6_hac_lac.m +++ b/matlab/uniaxial_6_hac_lac.m @@ -6,7 +6,6 @@ s = zpk('s'); %% Path for functions, data and scripts addpath('./mat/'); % Path for data -addpath('./src/'); % Path for functions %% Colors for the figures colors = colororder; diff --git a/nass-uniaxial-model.html b/nass-uniaxial-model.html index 3218a19..9f401cb 100644 --- a/nass-uniaxial-model.html +++ b/nass-uniaxial-model.html @@ -3,9 +3,9 @@ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> - + -Unidirectional Model - Report +NASS - Uniaxial Model @@ -34,49 +34,57 @@ | HOME
-

Unidirectional Model - Report

+

NASS - Uniaxial Model

@@ -84,89 +92,76 @@

This report is also available as a pdf.

-
-
(let ((org-export-before-parsing-hook (lambda (backend)
-                                        (let* ((data (org-element-parse-buffer))
-                                               (headline-labels (append
-                                                                 (org-element-map data 'headline
-                                                                   (lambda (hl)
-                                                                     (when (org-element-property :CUSTOM_ID hl)
-                                                                       (org-element-property :CUSTOM_ID hl))))
-                                                                 (org-element-map data 'link
-                                                                   (lambda (lnk)
-                                                                     (when
-                                                                         (and (string= "label" (org-element-property :type lnk))
-                                                                              (eq 'headline (car (org-element-property :parent lnk))))
-                                                                       (org-element-property :path lnk))))))
-                                               (table-labels (append
-                                                              (org-element-map data 'table
-                                                                (lambda (tbl)
-                                                                  (when (org-element-property :name tbl)
-                                                                    (org-element-property :name tbl))))
-                                                              (org-element-map data 'link
-                                                                (lambda (lnk)
-                                                                  (when (and (string= "label" (org-element-property :type lnk))
-                                                                             (eq 'table (car (org-element-property :parent lnk))))
-                                                                    (org-element-property :path lnk)))))))
-                                          ;; now we need to map over these and replace them
-                                          (mapc (lambda (label)
-                                                  (goto-char (point-min))
-                                                  (while (re-search-forward (format "ref:%s" label) nil t)
-                                                    (replace-match (format "[[%s]]" label))))
-                                                headline-labels)
-                                          (mapc (lambda (label)
-                                                  (goto-char (point-min))
-                                                  (while (re-search-forward (format "label:%s" label) nil t)
-                                                    (replace-match (format "<<%s>>" label))))
-                                                headline-labels)
-                                          (mapc (lambda (label)
-                                                  (goto-char (point-min))
-                                                  (while (re-search-forward (format "ref:%s" label) nil t)
-                                                    (replace-match (format "[[%s]]" label))))
-                                                table-labels)))))
+

+In this report, a uniaxial model of the Nano Active Stabilization System (NASS) is developed and used to have a first idea of the challenges involved in this complex system.
+Note that in this document, only the vertical direction is considered (which is the most stiff), but other directions were considered as well and yields similar conclusions.
+The model is schematically shown in Figure fig:uniaxial_overview_model_sections where the colors are representing the studied parts in different sections.
+

 
-  (org-open-file (org-html-export-to-html)))
-
+

+In order to have a relevant model, the micro-station dynamics is first identified and its model is tuned to match the measurements (Section sec:micro_station_model). +Then, a model of the nano-hexapod is added on top of the micro-station. +With added sample and sensors, this gives a uniaxial dynamical model of the NASS that will be used for further analysis (Section sec:nano_station_model). +

+ +

+The disturbances affecting the position accuracy are identified experimentally (Section sec:uniaxial_disturbances) and included in the model for dynamical noise budgeting (Section sec:uniaxial_noise_budgeting). +In all the following analysis, there nano-hexapod stiffnesses are considered to better understand the trade-offs and to find the most adequate stiffness. +Three sample masses are also considered to verify the robustness of the applied control strategies to a change of sample. +

+ +

+Three active damping techniques are then applied on the nano-hexapod. +This helps to reduce the effect of disturbances as well as render the system easier to control afterwards (Section sec:uniaxial_active_damping). +

+ +

+Once the system is well damped, a feedback position controller is applied, and the obtained performances are compared (Section sec:uniaxial_position_control). +

+ +

+Conclusion remarks are given in Section sec:conclusion. +

+ + +
+

uniaxial_overview_model_sections.png +

+

Figure 1: Uniaxial Micro-Station model in blue (Section sec:micro_station_model), Nano-Hexapod and sample models in red (Section sec:nano_station_model), Disturbances in yellow (Section sec:uniaxial_disturbances), Active Damping in green (Section sec:uniaxial_active_damping) and Position control in purple (Section sec:uniaxial_position_control)

-

-The report is organized as follow: -

-
    -
  • section sec:micro_station_model: an uniaxial of the micro-station is developped and compared with measurements.
    • -
  • section sec:nano_station_model: a model of the NASS is added on top of the model of the micro-station and some properties of the obtained model are studied.
    • -
  • section sec:noise_budget: after a definition of the control objective, an open loop noise budgeting is done. It consists of studying the effect of the sources of perturbation on the control objective. Some conclusions can be made on the open loop performances of both systems.
    • -
  • section sec:active_damping: active damping is applied on the system using integral force feedback in order to obtain critical damping of the NASS’s mode.
    • -
  • section sec:feedback_control: feedback control is applied using both actuators.
    • -
- -
-

1. Micro Station Model

+
+

1. Micro Station Model

- +

In this section, a uni-axial model of the micro-station is tuned in order to match measurements made on the micro-station +The measurement setup is shown in Figure fig:micro_station_first_meas_dynamics where several geophones are fixed to the micro-station and an instrumented hammer is used to inject forces on different stages of the micro-station. +

+ +

+From the measured frequency response functions (FRF), the model can be tuned to approximate the uniaxial dynamics of the micro-station.

-
+

micro_station_first_meas_dynamics.jpg

-

Figure 1: Experimental Setup for the first dynamical measurements on the Micro-Station. Geophones are fixed to the micro-station, and the granite as well as the micro-hexapod’s top platform are impact with an instrumented hammer

+

Figure 2: Experimental Setup for the first dynamical measurements on the Micro-Station. Geophones are fixed to the micro-station, and the granite as well as the micro-hexapod’s top platform are impact with an instrumented hammer

-
-

1.1. Measured dynamics

+
+

1.1. Measured dynamics

-The measurement setup is schematically shown in Figure 2 where: +The measurement setup is schematically shown in Figure fig:micro_station_meas_dynamics_schematic where:

    -
  • Two hammer hits are performed, one on the Granite, and one on the micro-hexapod’s top platform
    • -
  • The inertial motion of the granite and the micro-hexapod’s top platform are measured using geophones.
    • +
  • Two hammer hits are performed, one on the Granite (force \(F_g\)), and one on the micro-hexapod’s top platform (force \(F_h\))
    • +
  • The inertial motion of the granite \(x_g\) and the micro-hexapod’s top platform \(x_h\) are measured using geophones.

@@ -179,15 +174,14 @@ From the forces applied by the instrumented hammer and the responses of the geop -

+

micro_station_meas_dynamics_schematic.png

-

Figure 2: Measurement setup - Schematic

+

Figure 3: Measurement setup - Schematic

-

-Due to the bad coherence at low frequency, the frequency response functions are only shown between 20 and 200Hz (Figure 3). +Due to the bad coherence at low frequency, the frequency response functions are only shown between 20 and 200Hz (Figure fig:uniaxial_measured_frf_vertical).

@@ -197,19 +191,19 @@ load('meas_microstation_frf.mat');
-
-

measured_frf_vertical.png +

+

uniaxial_measured_frf_vertical.png

-

Figure 3: Measured Frequency Response Functions in the vertical direction

+

Figure 4: Measured Frequency Response Functions in the vertical direction

- -
-

1.3. Comparison of the model and measurements

+
+

1.3. Comparison of the model and measurements

-The comparison between the measurements and the model is done in Figure 5. +The comparison between the measurements and the model is done in Figure fig:uniaxial_comp_frf_meas_model.

-Ad the model is simplistic, the goal is not to match exactly the measurement but to have a first approximation. -More accurate models will be developed and tuned later on. +As the model is simplistic, the goal is not to match exactly the measurement but to have a first approximation. +More accurate models will be used later on.

-
-

comp_frf_meas_model.png +

+

uniaxial_comp_frf_meas_model.png

-

Figure 5: Comparison of the measured FRF and identified ones from the uni-axial model

+

Figure 6: Comparison of the measured FRF and identified ones from the uni-axial model

- -
-

2. Nano-Hexapod Model

+
+

2. Nano-Hexapod Model

- +

-A model of the nano-hexapod and sample is now added on top of the uni-axial model of the micro-station (Figure 6). +A model of the nano-hexapod and sample is now added on top of the uni-axial model of the micro-station (Figure fig:uniaxial_model_micro_station-nass).

@@ -298,15 +302,20 @@ Disturbances (shown in red) are: The control signal is the force applied by the nano-hexapod \(f\) and the measurement is the relative motion between the sample and the granite \(d\).

- -
-

uniaxial-model-micro-station-nass.png +

+The sample is here considered as a rigid body and rigidly fixed to the nano-hexapod. +The effect of having resonances between the sample’s point of interest and the nano-hexapod actuator will be considered in further analysis.

-

Figure 6: Uni-axial model of the micro-station with added nano-hexapod and sample

+ + +
+

uniaxial_model_micro_station-nass.png +

+

Figure 7: Uni-axial model of the micro-station with added nano-hexapod (represented in blue) and sample (represented in green)

-
-

2.1. Nano-Hexapod Parameters

+
+

2.1. Nano-Hexapod Parameters

The parameters for the nano-hexapod and sample are: @@ -334,39 +343,42 @@ ms = 10; %

-
-

2.2. Obtained Dynamics

+
+

2.2. Obtained Dynamics

-The sensitivity to disturbances (i.e. \(x_f\), \(f_t\) and \(f_s\)) are shown in Figure 7. -The plant (i.e. the transfer function from actuator force \(f\) to measured displacement \(d\)) is shown in Figure 8. +The sensitivity to disturbances (i.e. \(x_f\), \(f_t\) and \(f_s\)) are shown in Figure fig:uniaxial_sensitivity_dist_first_params. +The plant (i.e. the transfer function from actuator force \(f\) to measured displacement \(d\)) is shown in Figure fig:uniaxial_plant_first_params. +

+ +

+For further analysis, 9 configurations are considered: three nano-hexapod stiffnesses (\(k_n = 0.01\,N/\mu m\), \(k_n = 1\,N/\mu m\) and \(k_n = 100\,N/\mu m\)) combined with three sample’s masses (\(m_s = 1\,kg\), \(m_s = 25\,kg\) and \(m_s = 50\,kg\)).

-
+

uniaxial_sensitivity_dist_first_params.png

-

Figure 7: Sensitivity to disturbances

+

Figure 8: Sensitivity to disturbances

-
+

uniaxial_plant_first_params.png

-

Figure 8: Bode Plot of the transfer function from actuator forces to measured displacement by the metrology

+

Figure 9: Bode Plot of the transfer function from actuator forces to measured displacement by the metrology

- -
-

3. Disturbance Identification

+
+

3. Disturbance Identification

- +

-In order to measure disturbances, two geophones are used, on located on the floor and on on the micro-hexapod’s top platform (see Figure 9). +In order to measure disturbances, two geophones are used, on located on the floor and on on the micro-hexapod’s top platform (see Figure fig:micro_station_meas_disturbances).

@@ -374,18 +386,21 @@ The geophone on the floor is used to measured the floor motion \(x_f\) while the

-
+

micro_station_meas_disturbances.png

-

Figure 9: Disturbance measurement setup - Schematic

+

Figure 10: Disturbance measurement setup - Schematic

-

-The voltage generated by each geophone is amplified using a voltage amplifier (gain of 60dB) before going to the ADC. + +

+

micro_station_dynamical_id_setup.jpg

+

Figure 11: Two geophones are used to measure the micro-station vibrations induced by the scanning of the \(T_y\) and \(R_z\) stages

-
-

3.1. Ground Motion

+
+
+

3.1. Ground Motion

The geophone fixed to the floor to measure the floor motion. @@ -393,14 +408,16 @@ The geophone fixed to the floor to measure the floor motion. 

%% Load floor motion data
+% t: time in [s]
+% V: measured voltage genrated by the geophone and amplified by a 60dB gain voltage amplifier [V]
 load('ground_motion_measurement.mat', 't', 'V');

-The sensitivity of the geophone as well as the gain of the voltage amplifier are taken into account to reconstruct the floor displacement. +The voltage generated by each geophone is amplified using a voltage amplifier (gain of 60dB) before going to the ADC. +The sensitivity of the geophone as well as the gain of the voltage amplifier are then taken into account to reconstruct the floor displacement.

- 
%% Sensitivity of the geophone
 S0 = 88; % Sensitivity [V/(m/s)]
@@ -420,7 +437,6 @@ G_geo = 1/S/G0
 

 The PSD \(S_{V_f}\) of the measured voltage \(V_f\) is computed.
 

-
 

 
%% Compute measured voltage PSD
 Fs = 1/(t(2)-t(1)); % Sampling Frequency [Hz]
@@ -452,20 +468,20 @@ psd_xf = psd_V.*abs(squeeze(freqresp(G_geo, f,
 

 
 

-The amplitude spectral density \(\Gamma_{x_f}\) of the measured displacement \(x_f\) is shown in Figure 10.
+The amplitude spectral density \(\Gamma_{x_f}\) of the measured displacement \(x_f\) is shown in Figure fig:asd_floor_motion_id31.
 

 
 
-

+

 
asd_floor_motion_id31.png
 

-
Figure 10: Measured Amplitude Spectral Density of the Floor motion on ID31

+
Figure 12: Measured Amplitude Spectral Density of the Floor motion on ID31
-
-

3.2. Stage Vibration

+
+

3.2. Stage Vibration

During Spindle rotation (here at 6rpm), the granite velocity and micro-hexapod’s top platform velocity are measured with the geophones. @@ -473,7 +489,11 @@ During Spindle rotation (here at 6rpm), the granite velocity and micro-hexapod&r 

%% Measured velocity of granite and hexapod during spindle rotation
+% t: time in [s]
+% vg: measured granite velocity [m/s]
+% vg: measured micro-hexapod's top platform velocity [m/s]
 load('meas_spindle_on.mat', 't', 'vg', 'vh');
+spindle_off = load('meas_spindle_off.mat', 't', 'vg', 'vh'); % No Rotation
@@ -486,18 +506,20 @@ Fs = 1/(t(2)-t win = hanning(ceil(2*Fs)); % Hanning window [psd_vft, f] = pwelch(vh-vg, win, [], [], Fs); % [(m/s)^2/Hz] +[psd_off, ~] = pwelch(spindle_off.vh-spindle_off.vg, win, [], [], Fs); % [(m/s)^2/Hz]

-It is then integrated to obtain the Amplitude Spectral Density of the relative motion (Figure 11). +It is then integrated to obtain the Amplitude Spectral Density of the relative motion which is compared with a non-rotating case (Figure fig:asd_vibration_spindle_rotation). +It is shown that the spindle rotation induces vibrations in a wide frequency spectrum.

-
+

asd_vibration_spindle_rotation.png

-

Figure 11: Measured Amplitude Spectral Density of the relative motion between the granite and the micro-hexapod’s top platform during Spindle rotating

+

Figure 13: Measured Amplitude Spectral Density of the relative motion between the granite and the micro-hexapod’s top platform during Spindle rotating

@@ -515,37 +537,39 @@ with:

  • \(G_{\text{model}}\) the transfer function (extracted from the uniaxial model) from \(f_t\) to the relative motion between the micro-hexapod’s top platform and the granite
    • -
    -
    %% Power Spectral Density of the equivalent force ft
-psd_ft = (psd_vft./(2*pi*f).^2)./abs(squeeze(freqresp(G('Dh', 'ft') - G('Dg', 'ft'), f, 'Hz'))).^2;
-
    -
    -

    -The obtained amplitude spectral density of the disturbance force \(f_t\) is shown in Figure 12. +The obtained amplitude spectral density of the disturbance force \(f_t\) is shown in Figure fig:asd_disturbance_force.

    -
    +

    asd_disturbance_force.png

    -

    Figure 12: Estimated disturbance force ft from measurement and uniaxial model

    +

    Figure 14: Estimated disturbance force ft from measurement and uniaxial model

    + +

    +The vibrations induced by the \(T_y\) stage are not considered here as: +

    +
      +
    • the induced vibrations have less amplitude than the vibrations induced by the \(R_z\) stage
      • +
    • it can be scanned at lower velocities if the induced vibrations are an issue
      • +
    -
    -

    4. Open-Loop Dynamic Noise Budgeting

    +
    +

    4. Open-Loop Dynamic Noise Budgeting

    - +

    Now that we have a model of the NASS and an estimation of the power spectral density of the disturbances, it is possible to perform an open-loop dynamic noise budgeting.

    -
    -

    4.1. Sensitivity to disturbances

    +
    +

    4.1. Sensitivity to disturbances

    From the Uni-axial model, the transfer function from the disturbances (\(f_s\), \(x_f\) and \(f_t\)) to the displacement \(d\) are computed. @@ -561,12 +585,12 @@ This is done for two extreme sample masses \(m_s = 1\,\text{kg}\) and \(m

    -The obtained sensitivity to disturbances for the three nano-hexapod stiffnesses are shown in Figure 13 for the light sample (same conclusions can be drawn with the heavy one). +The obtained sensitivity to disturbances for the three nano-hexapod stiffnesses are shown in Figure fig:uniaxial_sensitivity_disturbances_nano_hexapod_stiffnesses for the light sample (same conclusions can be drawn with the heavy one).

    -
    +

    -From Figure 13, following can be concluded for the soft nano-hexapod: +From Figure fig:uniaxial_sensitivity_disturbances_nano_hexapod_stiffnesses, following can be concluded for the soft nano-hexapod:

    • It is more sensitive to forces applied on the sample (cable forces for instance), which is expected due to the lower stiffness
      • @@ -577,23 +601,23 @@ From Figure 13, following can be concluded for the
    -
    -

    sensitivity_disturbances_nano_hexpod_stiffnesses.png +

    +

    uniaxial_sensitivity_disturbances_nano_hexapod_stiffnesses.png

    -

    Figure 13: Sensitivity to disturbances for three different nano-hexpod stiffnesses

    +

    Figure 15: Sensitivity to disturbances for three different nano-hexpod stiffnesses

    -
    -

    4.2. Open-Loop Dynamic Noise Budgeting

    +
    +

    4.2. Open-Loop Dynamic Noise Budgeting

    Now, the power spectral density of the disturbances is taken into account to estimate the residual motion \(d\) in each case.

    -The Cumulative Amplitude Spectrum of the relative motion \(d\) due to both the floor motion \(x_f\) and the stage vibrations \(f_t\) are shown in Figure 14 for the three nano-hexapod stiffnesses. +The Cumulative Amplitude Spectrum of the relative motion \(d\) due to both the floor motion \(x_f\) and the stage vibrations \(f_t\) are shown in Figure fig:uniaxial_cas_d_disturbances_stiffnesses for the three nano-hexapod stiffnesses.

    @@ -601,25 +625,30 @@ It is shown that the effect of the floor motion is much less than the stage vibr

    -
    -

    cas_d_disturbances_stiffnesses.png +

    +

    uniaxial_cas_d_disturbances_stiffnesses.png

    -

    Figure 14: Cumulative Amplitude Spectrum of the relative motion d, due to both the floor motion and the stage vibrations (light sample)

    +

    Figure 16: Cumulative Amplitude Spectrum of the relative motion d, due to both the floor motion and the stage vibrations (light sample)

    -The total cumulative amplitude spectrum for the three nano-hexapod stiffnesses and for the two sample’s masses are shown in Figure 15. +The total cumulative amplitude spectrum for the three nano-hexapod stiffnesses and for the two sample’s masses are shown in Figure fig:uniaxial_cas_d_disturbances_payload_masses. The conclusion is that the sample’s mass has little effect on the cumulative amplitude spectrum of the relative motion \(d\).

    -
    -

    cas_d_disturbances_payload_masses.png +

    +

    uniaxial_cas_d_disturbances_payload_masses.png

    -

    Figure 15: Cumulative Amplitude Spectrum of the relative motion d due to all disturbances, for two sample masses

    +

    Figure 17: Cumulative Amplitude Spectrum of the relative motion d due to all disturbances, for two sample masses

    +
    +
    -
    +
    +

    4.3. Conclusion

    +
    +

    The conclusion is that in order to have a closed-loop residual vibration \(d \approx 20\,nm\text{ rms}\), if a simple feedback controller is used, the required closed-loop bandwidth would be:

    @@ -629,157 +658,543 @@ The conclusion is that in order to have a closed-loop residual vibration \(d \ap
  • \(\approx 100\,\text{Hz}\) for the stiff nano-hexapod (\(k_n = 100\,N/\mu m\))
    • +

    +This can be explain by the fact that above the suspension mode of the nano-hexapod, the stage vibrations are filtered out (see Figure fig:uniaxial_sensitivity_disturbances_nano_hexapod_stiffnesses). +

    + +

    +This gives a first advantage to having a soft nano-hexapod. +

    +
    -
    -

    5. Active Damping

    +
    +

    5. Active Damping

    - + +

    +

    +In this section, three active damping are applied on the nano-hexapod (see Figure fig:uniaxial_active_damping_strategies): Integral Force Feedback (IFF), Relative Damping Control (RDC) and Direct Velocity Feedback (DVF).

    -
    +

    +These active damping techniques are compared in terms of: +

    +
      +
    • Reduction of the effect of disturbances (i.e. \(x_f\), \(f_t\) and \(f_s\)) on the displacement \(d\)
      • +
    • Achievable damping
      • +
    • Robustness to a change of sample’s mass
      • +
    + + +

    uniaxial_active_damping_strategies.png

    -

    Figure 16: Three active damping strategies: Direct Velocity Feedback (DVF) using a geophone, Integral Force Feedback (IFF) using a force sensor, and Relative Motion Control (RMC) using a relative displacement sensor

    +

    Figure 18: Three active damping strategies: Integral Force Feedback (IFF) using a force sensor, Relative Damping Control (RDC) using a relative displacement sensor, and Direct Velocity Feedback (DVF) using a geophone

    -
    -

    5.1. Plant Dynamics

    +
    +

    5.1. Active Damping Strategies

    -The plant dynamics for all three active damping techniques are shown in Figure 17. +The Integral Force Feedback strategy consists of using a force sensor in series with the actuator (see Figure fig:uniaxial_active_damping_iff_equiv, left).

    -All have alternating poles and zeros. +The control strategy consists of integrating the measured force and feeding it back to the actuator: +

    +\begin{equation} +K_{\text{IFF}}(s) = \frac{g}{s} +\end{equation} + +

    +The mechanical equivalent of this strategy is to add a dashpot in series with the actuator stiffness with a damping coefficient equal to the stiffness of the actuator divided by the controller gain \(k/g\) (see Figure fig:uniaxial_active_damping_iff_equiv, right).

    -
    -

    uniaxial_plant_active_damping_techniques.png +

    +

    uniaxial_active_damping_iff_equiv.png

    -

    Figure 17: Plant dynamics for the three active damping techniques

    +

    Figure 19: Integral Force Feedback is equivalent as to add a damper in series with the stiffness (the initial damping is here neglected for simplicity)

    +
    + + +

    +For the Relative Damping Control strategy, a relative motion sensor that measures the motion of the actuator is used (see Figure fig:uniaxial_active_damping_rdc_equiv, left). +

    + +

    +The derivative of this relative motion is used for the feedback signal: +

    +\begin{equation} +K_{\text{RDC}}(s) = - g \cdot s +\end{equation} + +

    +The mechanical equivalent is to add a dashpot in parallel with the actuator with a damping coefficient equal to the controller gain \(g\) (see Figure fig:uniaxial_active_damping_rdc_equiv, right). +

    + + +
    +

    uniaxial_active_damping_rdc_equiv.png +

    +

    Figure 20: Relative Damping Control is equivalent as adding a damper in parallel with the actuator/relative motion sensor

    +
    + +

    +Finally, the Direct Velocity Feedback strategy consists of using an inertial sensor (usually a geophone), that measured the “absolute” velocity of the body fixed on top of the actuator (se Figure fig:uniaxial_active_damping_dvf_equiv, left). +

    + +

    +The measured velocity is then fed back to the actuator: +

    +\begin{equation} +K_{\text{DVF}}(s) = - g +\end{equation} + +

    +This is equivalent as to fix a dashpot (with a damping coefficient equal to the controller gain \(g\)) between the body (one which the inertial sensor is fixed) and an inertial reference frame (see Figure fig:uniaxial_active_damping_dvf_equiv, right). +This is usually refers to as “sky hook damper”. +

    + + +
    +

    uniaxial_active_damping_dvf_equiv.png +

    +

    Figure 21: Direct velocity Feedback using an inertial sensor is equivalent to a “sky hook damper”

    -
    -

    5.2. Achievable Damping

    +
    +

    5.2. Plant Dynamics for Active Damping

    -Controller used for the three active damping techniques are: +The plant dynamics for all three active damping techniques are shown in Figure fig:uniaxial_plant_active_damping_techniques. +All have alternating poles and zeros meaning that the phase is bounded to \(\pm 90\,\text{deg}\) which makes the controller very robust.

    -\begin{align} -K_{\text{IFF}}(s) &= \frac{g_{\text{IFF}}}{s} \\ -K_{\text{RMC}}(s) &= g_{\text{RMC}} s \\ -K_{\text{DVF}}(s) &= g_{\text{DVF}} -\end{align}

    -It is shown in the Root Locus in Figure 18 that all three active damping approach can lead to critical damping. +When the nano-hexapod’s suspension modes are at lower frequencies than the resonances of the micro-station (blue and red curves in Figure fig:uniaxial_plant_active_damping_techniques), the resonances of the micro-stations have little impact on the transfer functions from IFF and DVF. +

    + +

    +For the stiff nano-hexapod, the micro-station dynamics can be seen on the transfer functions in Figure fig:uniaxial_plant_active_damping_techniques. +Therefore, it is expected that the micro-station dynamics might impact the achievable damping if a stiff nano-hexapod is used.

    -
    -

    uniaxial_root_locus_damping_techniques.png +

    +

    uniaxial_plant_active_damping_techniques.png

    -

    Figure 18: Root Locus for the three damping techniques

    +

    Figure 22: Plant dynamics for the three active damping techniques (IFF: right, RDC: middle, DVF: left), for three nano-hexapod stiffnesses (\(k_n = 0.01\,N/\mu m\) in blue, \(k_n = 1\,N/\mu m\) in red and \(k_n = 100\,N/\mu m\) in yellow) and three sample’s masses (\(m_s = 1\,kg\): solid curves, \(m_s = 25\,kg\): dot-dashed curves, and \(m_s = 50\,kg\): dashed curves).

    -
    -

    5.3. Change of sensitivity to disturbances

    +
    +

    5.3. Achievable Damping - Root Locus

    -The sensitivity to disturbances (\(f_s\), \(f_t\) and \(x_f\)) for all three active damping techniques are compared in Figure 19. +The Root Locus are computed for the three nano-hexapod stiffnesses and for the three active damping techniques. +They are shown in Figure fig:uniaxial_root_locus_damping_techniques.

    -

    -Conclusions from Figure 19 are: +All three active damping approach can lead to critical damping of the nano-hexapod suspension mode. +

    + +

    +There is even a little bit of authority on micro-station modes with IFF and DVF applied on the stiff nano-hexapod (Figure fig:uniaxial_root_locus_damping_techniques, right) and with RDC for a soft nano-hexapod (Figure fig:uniaxial_root_locus_damping_techniques_micro_station_mode). +This can be explained by the fact that above the suspension mode of the soft nano-hexapod, the relative motion sensor acts as an inertial sensor for the micro-station top platform. Therefore, it is like DVF was applied (the nano-hexapod acts as a geophone!). +

    + + +
    +

    uniaxial_root_locus_damping_techniques.png +

    +

    Figure 23: Root Loci for the three active damping techniques (IFF in blue, RDC in red and DVF in yellow). This is shown for three nano-hexapod stiffnesses.

    +
    + + +
    +

    uniaxial_root_locus_damping_techniques_micro_station_mode.png +

    +

    Figure 24: Root Locus for the three damping techniques. It is shown that the RDC active damping technique has some authority on one mode of the micro-station. This mode corresponds to the suspension mode of the micro-hexapod.

    +
    +
    +
    + +
    +

    5.4. Change of sensitivity to disturbances

    +
    +

    +The sensitivity to disturbances (direct forces \(f_s\), stage vibrations \(f_t\) and floor motion \(x_f\)) for all three active damping techniques are compared in Figure fig:uniaxial_sensitivity_dist_active_damping. +The comparison is done with the nano-hexapod having a stiffness \(k_n = 1\,N/\mu m\). +

    + +
    +

    +Conclusions from Figure fig:uniaxial_sensitivity_dist_active_damping are:

    • IFF degrades the sensitivity to direct forces on the sample (i.e. the compliance) below the resonance of the nano-hexapod
      • -
    • RMC degrades the sensitivity to stage vibrations around the nano-hexapod’s resonance as compared to the other two methods
      • +
    • RDC degrades the sensitivity to stage vibrations around the nano-hexapod’s resonance as compared to the other two methods
    • both IFF and DVF degrades the sensitivity to floor motion below the resonance of the nano-hexapod
    -
    +

    uniaxial_sensitivity_dist_active_damping.png

    -

    Figure 19: Change of sensitivity to disturbance with all three active damping strategies

    +

    Figure 25: Change of sensitivity to disturbance with all three active damping strategies

    -
    -

    5.4. Noise Budgeting after Active Damping

    -
    +
    +

    5.5. Noise Budgeting after Active Damping

    +

    -Cumulative Amplitude Spectrum of the distance \(d\) with all three active damping techniques are compared in Figure 20. -All three active damping methods are giving similar results. +Cumulative Amplitude Spectrum of the distance \(d\) with all three active damping techniques are compared in Figure fig:uniaxial_cas_active_damping. +All three active damping methods are giving similar results (except the RDC which is a little bit worse for the stiff nano-hexapod).

    -Active Damping helps to lower the vibrations and less bandwidth is required for the position controller. +Compare to the open-loop case, the active damping helps to lower the vibrations induced by the nano-hexapod resonance.

    -
    +

    uniaxial_cas_active_damping.png

    -

    Figure 20: Cumulative Amplitude Spectrum of the distance \(d\) with all three active damping techniques

    -
    +

    Figure 26: Comparison of the cumulative amplitude spectrum (CAS) of the distance \(d\) for all three active damping techniques (OL in black, IFF in blue, RDC in red and DVF in yellow).

    -
    -

    6. Position Feedback Controller

    -
    +
    +

    5.6. Obtained Damped Plant

    +

    - +The transfer functions from the plant input \(f\) to the relative displacement \(d\) while the active damping is implemented are shown in Figure fig:uniaxial_damped_plant_three_active_damping_techniques. +All three active damping techniques yield similar damped plants. +

    + + +
    +

    uniaxial_damped_plant_three_active_damping_techniques.png +

    +

    Figure 27: Obtained damped transfer function from f to d for the three damping techniques

    +
    + +

    +The damped plants are shown in Figure fig:uniaxial_damped_plant_change_sample_mass for all three techniques, with the three considered nano-hexapod stiffnesses and sample’s masses. +

    + +
    +

    uniaxial_damped_plant_change_sample_mass.png +

    +

    Figure 28: Damped plant \(d/f\) - Robustness to change of sample’s mass for all three active damping techniques. Grey curves are the open-loop (i.e. undamped) plants.

    +
    +
    +
    + +
    +

    5.7. Robustness to change of payload’s mass

    +
    +

    +The Root Locus for the three damping techniques are shown in Figure fig:uniaxial_active_damping_robustness_mass_root_locus for three sample’s mass (1kg, 25kg and 50kg). +The closed-loop poles are shown by the squares for a specific gain. +

    + +

    +We can see that having heavier samples yields larger damping for IFF and smaller damping for RDC and DVF. +

    + + +
    +

    uniaxial_active_damping_robustness_mass_root_locus.png +

    +

    Figure 29: Active Damping Robustness to change of sample’s mass - Root Locus for all three damping techniques with 3 different sample’s masses

    +
    +
    +
    + +
    +

    5.8. Conclusion

    +
    +
    +

    +Conclusions for Active Damping:

      -
    • [X] Explain the HAC-LAC control strategy
      • -
    • [ ] Show the plant for the three stiffnesses
      • -
    • [ ] Estimate the required bandwidth to attain the specifications
      • +
    • All three active damping techniques yields good damping (Figure fig:uniaxial_root_locus_damping_techniques) and similar remaining vibrations (Figure fig:uniaxial_cas_active_damping)
      • +
    • The obtained damped plants (Figure fig:uniaxial_damped_plant_change_sample_mass) are equivalent for the three active damping techniques
      • +
    • Which one to be used will be determined with the use of more accurate models and will depend on which is the easiest to implement in practice
    +
    + + + + +++ ++ ++ ++ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    Table 1: Comparison of active damping strategies
     IFFRDCDVF
    SensorForce sensorRelative motion sensorInertial sensor
    DampingUp to criticalUp to criticalUp to Critical
    RobustnessRequires collocationRequires collocationImpacted by geophone resonances
    \(f_s\) Disturbance\(\nearrow\) at low frequency\(\searrow\) near resonance\(\searrow\) near resonance
    \(f_t\) Disturbance\(\searrow\) near resonance\(\nearrow\) near resonance\(\searrow\) near resonance
    \(x_f\) Disturbance\(\nearrow\) at low frequency\(\searrow\) near resonance\(\nearrow\) at low frequency
    +
    +
    +
    + +
    +

    6. Position Feedback Controller

    +

    -The High Authority Control - Low Authority Control (HAC-LAC) architecture is shown in Figure 21. + +

    +

    +The High Authority Control - Low Authority Control (HAC-LAC) architecture is shown in Figure fig:uniaxial_hac_lac_architecture. It corresponds to a two step control strategy:

    • First, an active damping controller \(\bm{K}_{\textsc{LAC}}\) is implemented (see Section sec:uniaxial_active_damping). It allows to reduce the vibration level, and it also makes the damped plant (transfer function from \(u^{\prime}\) to \(y\)) easier to control than the undamped plant (transfer function from \(u\) to \(y\)).
      • -
    • Then, a position controller \(K_{\textsc{HAC}}\) is implemented.
      • +
    • Then, a position controller \(\bm{K}_{\textsc{HAC}}\) is implemented.
      • +
    + +

    +Combined with the uniaxial model, it is shown in Figure fig:uniaxial_hac_lac_model. +

    + +
    +
    + +
    +

    uniaxial_hac_lac_architecture.png +

    +
    + +
    +
    + +
    +

    uniaxial_hac_lac_model.png +

    +
    + +
    + +
    +
    +
    +

    6.1. Damped Plant Dynamics

    +
    +

    +As was shown in Section sec:uniaxial_active_damping, all three proposed active damping techniques yield similar damping plants. +Therefore, Integral Force Feedback will be used in this section to study the HAC-LAC performances. +

    + +

    +The obtained damped plants for the three nano-hexapod stiffnesses are shown in Figure fig:uniaxial_hac_iff_damped_plants_masses. +

    + + +
    +

    uniaxial_hac_iff_damped_plants_masses.png +

    +

    Figure 30: Obtained damped plant using Integral Force Feedback for three sample’s masses

    +
    +
    +
    + +
    +

    6.2. Position Feedback Controller

    +
    +

    +The objective now is to design a position feedback controller for each of the three nano-hexapods that are robust to the change of sample’s mass. +

    + +

    +The required feedback bandwidth was approximately determined un Section sec:uniaxial_noise_budgeting: +

    +
      +
    • \(\approx 10\,\text{Hz}\) for the soft nano-hexapod (\(k_n = 0.01\,N/\mu m\)). +Near this frequency, the plants are equivalent to a mass line. +The gain of the mass line can vary up to a fact \(\approx 5\) (suspended mass from \(16\,kg\) up to \(65\,kg\)). +This mean that the designed controller will need to have large gain margins to be robust to the change of sample’s mass.
      • +
    • \(\approx 50\,\text{Hz}\) for the relatively stiff nano-hexapod (\(k_n = 1\,N/\mu m\)). +Similarly to the soft nano-hexapod, the plants near the crossover frequency are equivalent to a mass line. +It will be probably easier to have a little bit more bandwidth in this configuration to be further away from the nano-hexapod suspension mode.
      • +
    • \(\approx 100\,\text{Hz}\) for the stiff nano-hexapod (\(k_n = 100\,N/\mu m\)). +Contrary to the two first nano-hexapod stiffnesses, here the plants have more complex dynamics near the wanted crossover frequency. +The micro-station is not stiff enough to have a clear stiffness line at this frequency. +Therefore, there are both a change of phase and gain depending on the sample’s mass. +This makes the robust design of the controller a little bit more complicated.
    - -
    -

    uniaxial_hac_lac_architecture.png +

    +Position feedback controllers are designed for each nano-hexapod such that it is stable for all considered sample masses with similar stability margins (see Nyquist plots in Figure fig:uniaxial_nyquist_hac). +These high authority controllers are generally composed of a two integrators at low frequency for disturbance rejection, a lead to increase the phase margin near the crossover frequency and a low pass filter to increase the robustness to high frequency dynamics. +The loop gains for the three nano-hexapod are shown in Figure fig:uniaxial_loop_gain_hac. +We can see that:

    -

    Figure 21: High Authority Control - Low Authority Control (HAC-LAC) architecture

    +
      +
    • for the soft and moderately stiff nano-hexapod, the crossover frequency varies a lot with the sample mass. +This is due to the fact that the crossover frequency corresponds to the mass line of the plant.
      • +
    • for the stiff nano-hexapod, the obtained crossover frequency is not at high as what was estimated necessary. +The crossover frequency in that case is close to the stiffness line of the plant, which makes the robust design of the controller easier.
      • +
    + +

    +Note that these controller were quickly tuned by hand and not designed using any optimization methods. +The goal is just to have a first estimation of the attainable performances. +

    + + +
    +

    uniaxial_nyquist_hac.png +

    +

    Figure 31: Nyquist Plot - Hight Authority Controller for all three nano-hexapod stiffnesses (soft one in blue, moderately stiff in red and very stiff in yellow) and all sample masses (corresponding to the three curves of each color)

    +
    + + +
    +

    uniaxial_loop_gain_hac.png +

    +

    Figure 32: Loop Gain - High Authority Controller for all three nano-hexapod stiffnesses (soft one in blue, moderately stiff in red and very stiff in yellow) and all sample masses (corresponding to the three curves of each color)

    -
    -

    6.1. Plant Dynamics

    +
    + +
    +

    6.3. Closed-Loop Noise Budgeting

    +
    +

    +The high authority position feedback controllers are then implemented and the closed-loop sensitivity to disturbances are computed. +These are compared with the open-loop and damped plants cases in Figure fig:uniaxial_sensitivity_dist_hac_lac for just one configuration (moderately stiff nano-hexapod with 25kg sample’s mass). +As expected, the sensitivity to disturbances is decreased in the controller bandwidth and slightly increase outside this bandwidth. +

    + + +
    +

    uniaxial_sensitivity_dist_hac_lac.png +

    +

    Figure 33: Change of sensitivity to disturbances with LAC and with HAC-LAC

    +
    + +

    +The cumulative amplitude spectrum of the motion \(d\) is computed for all nano-hexapod configurations, all sample masses and in the open-loop (OL), damped (IFF) and position controlled (HAC-IFF) cases. +The results are shown in Figure fig:uniaxial_cas_hac_lac. +Obtained root mean square values of the distance \(d\) are better for the soft nano-hexapod (\(\approx 25\,nm\) to \(\approx 35\,nm\) depending on the sample’s mass) than for the stiffer nano-hexapod (from \(\approx 30\,nm\) to \(\approx 70\,nm\)). +

    + + +
    +

    uniaxial_cas_hac_lac.png +

    +

    Figure 34: Cumulative Amplitude Spectrum for all three nano-hexapod stiffnesses - Comparison of OL, IFF and HAC-LAC cases

    +
    +
    +
    +
    + +
    +

    7. Conclusion

    +
    +

    + +

    + +

    +In this study, a uniaxial model of the nano-active-stabilization-system has been tuned both from dynamical measurements (Section sec:micro_station_model) and from disturbances measurements (Section sec:uniaxial_disturbances). +

    + +

    +It has been shown that three active damping techniques can be used to critically damp the nano-hexapod resonances (Section sec:uniaxial_active_damping). +However, this model does not allows to determine which one is most suited to this application. +

    + +

    +Finally, position feedback controllers have been developed for three considered nano-hexapod stiffnesses. +These controllers were shown to be robust to the change of sample’s masses, and to provide good rejection of disturbances. +It has been found that having a soft nano-hexapod makes the plant dynamics easier to control (because decoupled from the micro-station dynamics) and requires less position feedback bandwidth to fulfill the requirements. +The moderately stiff nano-hexapod (\(k_n = 1\,N/\mu m\)) is requiring a bit more position feedback bandwidth, but it still seems to give acceptable results. +However, the stiff nano-hexapod is the most complex to control and gives the worst positioning performances. +

    Author: Dehaeze Thomas

    -

    Created: 2023-02-14 Tue 13:19

    +

    Created: 2023-02-17 Fri 11:29

    diff --git a/nass-uniaxial-model.matlab b/nass-uniaxial-model.matlab index 4a10d89..71d1ef3 100644 --- a/nass-uniaxial-model.matlab +++ b/nass-uniaxial-model.matlab @@ -3,7 +3,6 @@ %% Path for functions, data and scripts addpath('./matlab/mat/'); % Path for data -addpath('./matlab/src/'); % Path for functions addpath('./matlab/'); % Path for scripts @@ -12,7 +11,6 @@ addpath('./matlab/'); % Path for scripts %% Path for functions, data and scripts addpath('./mat/'); % Path for data -addpath('./src/'); % Path for functions % Initialize other elements % #+NAME: m-init-other diff --git a/nass-uniaxial-model.org b/nass-uniaxial-model.org index f51cdb3..ba309aa 100644 --- a/nass-uniaxial-model.org +++ b/nass-uniaxial-model.org @@ -4174,7 +4174,6 @@ However, the stiff nano-hexapod is the most complex to control and gives the wor #+BEGIN_SRC matlab %% Path for functions, data and scripts addpath('./matlab/mat/'); % Path for data -addpath('./matlab/src/'); % Path for functions addpath('./matlab/'); % Path for scripts #+END_SRC @@ -4182,7 +4181,6 @@ addpath('./matlab/'); % Path for scripts #+BEGIN_SRC matlab %% Path for functions, data and scripts addpath('./mat/'); % Path for data -addpath('./src/'); % Path for functions #+END_SRC ** Initialize other elements